Distributed and parallel computing in MADM domain using the OPTCHOICE software

  • Authors:

  • Cornel Resteanu;Marius Somodi;Marin Andreica;Electra Mitan

  • Affiliations:

  • National Institute for Research and Development in Informatics, Bucharest, Romania;Department of Mathematics, University of Northern Iowa, Cedar Falls, IA;Economic Studies Academy, Bucharest, Romania;National Institute for Research and Development in Informatics, Bucharest, Romania

  • Venue:
  • ACS'07 Proceedings of the 7th Conference on 7th WSEAS International Conference on Applied Computer Science - Volume 7
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

The paper presents a method for solving the general Multi-Attribute Decision Making (MADM) problems, by distributed and parallel computing, with the OPTCHOICE software. One presents the scheduling and load balancing algorithm for concurrent solving of sets of such problems on a given number of parallel computers. An analysis on the construction of a MADM problem is made; in this way, a decomposition tree having the decision-makers on the first level, the states of nature on the second level, and the attributes of the problem on the third level is emphasized. Corroborated with the analysis of the problem's data, the above results conduct at the conviction that a parallel algorithm for solving the MADM general problem, starting from the MADM particular problem, is possible. At each tree's level one can state independent particular sub-problems that are solved in parallel, the sub-problems at a superior level waiting for the solutions of the sub-problems at the current level. Finally, the classical TOPSIS method is presented running in the parallel and multi-level context.